Research Article | Open Access
The Heterogeneous Effects of FDI and Foreign Trade on CO2 Emissions: Evidence from China
The environmental impacts of foreign direct investment (FDI) and foreign trade have attracted much attention recently. This paper employs panel quantile regression to explore the effects of FDI and foreign trade on Chinese provincial CO2 emissions for the period of 1997-2014. The results indicate that the effect of FDI on CO2 emissions is negative and significant except at the and quantiles. Foreign trade has a significant negative effect on CO2 emissions at upper quantiles, and the degree of the effect increases gradually with the increase of CO2 emissions. The results also suggest that the inverted U-shaped environmental Kuznets curve (EKC) is valid only in the least and most polluted provinces. Nevertheless, the positive indirect effects of FDI and foreign trade on CO2 emissions are greater than the negative direct effects; thus the total effects are positive. Finally, several policy implications are proposed for China based on the empirical results obtained.
Since the Reform and Opening in 1978, China has witnessed and is still witnessing high speed economic development with an annual growth rate that exceeds 10%, ranking second throughout the world . As the most important engine of economic development, FDI and foreign trade have made tremendous contributions to the rapid growth of economy. However, the inflow of FDI and the rising level of foreign trade are also identified to be fundamental causes of pollution in the environment . Already the world’s largest carbon emitter, China emits 9.15 billion tons of carbon emissions, accounting for 27.3% of the global total emissions in 2015 . Such status quo has raised widespread public concern about the unprecedented rise in CO2 emissions, which placed great pressure to China in dealing with anthropogenic environmental impact. Presently, economic globalization and trade liberalization have greatly promoted the inflow of FDI and the rapid growth of foreign trade. Assessing the environmental consequences caused by FDI and foreign trade will be of important significance in both restraining CO2 emissions and achieving sustainable development.
Previous studies on the environmental impacts of FDI and foreign trade mainly concentrated on the ‘direct’ mechanism, while little attention has been paid to the ‘indirect’ mechanism that FDI and foreign trade affecting CO2 emissions through their impacts on economic development . According to the environmental Kuznets curve (EKC) hypothesis, the nexus between GDP growth and emissions is nonlinear and shaped like the typical inverted U . The argument in favor of this hypothesis postulates that environmental pollution initially increases with the increase of per capita income but begins to decrease rapidly after the per capita income exceeds the threshold point. The applicability of EKC hypothesis has been widely verified in various countries and time periods. For example, in a study of 25 OECD countries, Jebli et al.  found an inverted U-shaped relationship between per capita carbon emissions and per capita real GDP over the period of 1980 to 2010. Likewise, the investigation of Riti et al.  supported the existence of EKC hypothesis in China for the period from 1970 to 2015. Moreover, many studies have highlighted the important role of FDI and foreign trade in accelerating economic growth. Li and Liu  documented that FDI can significantly promote economic growth in a global scale of 84 countries. Lau et al.  claimed that foreign trade is positively correlated with economic growth. This raises the question that, besides the direct effects, FDI and foreign trade may indirectly affect CO2 emissions by promoting economic growth. The neglect of the indirect effects may lead to the underestimation or overestimation to the environmental influences of FDI and foreign trade. Hence, further explorations of the indirect effects of FDI and foreign trade on environment are warranted to prevent their real impacts from being misestimated.
Besides, studies have ignored the likelihood that the impacts of FDI and foreign trade on CO2 emissions might present heterogeneity throughhout the emissions distribution. Most previous literatures concerned with the determinants of CO2 emissions adopted the traditional ordinary least square (OLS) method. Cade and Noon  argued that OLS estimations might be seriously under- or overestimated in heterogeneous distribution. There are pronounced differences in CO2 emissions among Chinese provinces. The effects of driving factors on CO2 emissions may be affected by different levels of carbon emissions; that is, the impact of the same determinant may be inconsistent in low- and high-emissions provinces. Since the widespread used OLS method can only estimate the mean effect, it is no longer reasonable and applicable. In the present article, this issue will be effectively solved by employing quantile regression approach. You et al.  argued that quantile regression can evaluate the effects of influence factors on CO2 emissions throughout the conditional distribution. Lin and Xu  used the same methodology and confirmed that the impacts of influencing factors on CO2 emissions in the manufacturing industry differ significantly across quantile levels.
This research extends existing literature in the following ways. First of all, this study quantitatively estimates the direct and indirect impacts of FDI and foreign trade on CO2 emissions in 30 Chinese provinces. At the same time taking into account the direct and indirect impacts can help us get more accurate information about how FDI and foreign trade influence CO2 emissions. Second, this paper employs panel quantile regression to study the environmental effects of FDI and foreign trade throughout the entire distribution of CO2 emissions. In contrast to the OLS method, quantile regression allows examining whether FDI and foreign trade consistently affect CO2 emissions across quantile levels, particularly in the lowest and highest quantiles. Third, certain related control variables are introduced into our model within the theoretical framework of STIRPAT (the stochastic impacts by regression on population, affluence, and technology), which can resolve the omitted-variable biases .
The remainder of the study proceeds as follows. Section 2 presents the relevant literature. Section 3 presents the empirical model, methodology, and data. The empirical results and relevant discussions are provided in Section 4. Section 5 reports the conclusions and policy suggestions.
2. Literature Review
2.1. FDI and CO2 Emissions
The majority of research on FDI–emissions linkage is aimed at verifying the existence of pollution haven hypothesis (PHH). According to the PHH, foreign firms in dirty sectors are more likely to relocate pollution activities from developed countries to poorly regulated developing countries to avoid domestic environmental control costs, which directly undermines the environmental interests of recipient countries . Many scholars have presented evidence in favor of this hypothesis. Among them, Shahbaz et al.  utilized the Fully Modified Ordinary Least Squares (FMOLS) method to analyze the nonlinear relation between FDI and carbon emissions in countries with different income levels. The results indicated that the influx of FDI would lead to the increase in carbon emissions. Tang and Tan  tested the correlation among FDI, carbon emissions, energy consumption, and income applying the cointegration test and Granger causality. Their results presented bidirectional causal connection between FDI and carbon emissions in Vietnam. Based on the 1981-2010 panel data of five ASEAN countries, Baek  introduced dynamic panel data model to explore the linkage among CO2 emissions, FDI, economic growth, and energy consumption. They came to the same conclusion that FDI is detrimental to the reduction of CO2 emissions. Seker et al.  employed the Autoregressive Distributed Lag (ARDL) model and Error Correction Model (ECM) to determine the effect of FDI on Turkish carbon emissions. The findings manifested that the elasticity of CO2 emissions to FDI is positive, either in the long-run or short-run. Solarin et al.  and Sun et al.  used the same approach to examine the existence of PHH in Ghana and China respectively. Their results illustrated that FDI is positively correlated with CO2 emissions, thus supporting the PHH. In the case of South and Southeast Asian (SSEA) countries, Behera and Dash  analyzed the impacts of FDI, urbanization, and energy consumption on CO2 emissions by means of the FMOLS and Dynamic Ordinary Least Square (DOLS) techniques. The results revealed that the influence of FDI on CO2 emissions depends on income level. Along with this, they found that FDI inflows deteriorate environment in countries with middle- and high-income. More recently, Salahuddin et al.  adopted the ARDL bounds testing approach to estimate the association among FDI, economic growth, financial development, electricity consumption, and carbon emissions in case of Kuwait. They also found that, in both the short and long run, FDI causes a rise in carbon emissions.
However, some studies have documented that FDI inflows are conductive to improve energy efficiency and ultimately curbing pollutant emissions, because FDI has positive spillover effects such as advanced management practices, up-to-date technologies, and employment expansion. This is also known as the pollution halo hypothesis. Evidence has provided strong support for this hypothesis. For example, Zhou et al.  used dynamic panel data model to assess the environmental impact of industrial structural transformation. The results stated that a larger amount of FDI inflowing would cut emissions, which validates the pollution halo effects in China. Mert and Bölük  used panel ARDL approach to estimate the effects of FDI and renewable energy consumption on environmental quality and found that FDI inhibits CO2 emissions in the 21 Kyoto countries. Moreover, in a study by Al-mulali and Tang , the validity of PHH was tested by FMOLS method. It was discovered that the increase of FDI has no impact on pollutant emissions in the Gulf Cooperation Council (GCC) countries.
2.2. Foreign Trade and CO2 Emissions
In addition to FDI, foreign trade may be another source of environmental pollution. The seminal exploration of trade-emissions nexus was provided by Grossman and Krueger , who concluded that the environment impact of trade can be divided into scale, technique, and composition effects. Since then a growing literature has been carried out to evaluate the impact of foreign trade on environmental quality, but empirical findings are ambiguous and fail to reach a consensus [27–29]. For instance, Tiwari et al.  inspected the relationship between economic growth, trade openness, CO2 emissions, and coal consumption in India by applying OLS method. The results reported that increasing trade openness would contribute to CO2 emissions. In a study by Ozturk and Al-mulali , the Generalized Method of Moments (GMM) and Two-Stage Least Squares (TSLS) method were adopted and the conclusions indicated that carbon emissions increase significantly with the increase of trade volume in Cambodia. On a regional level, Wang and Zhao  conducted an analysis of the driving factors that affect CO2 emissions in three different income regions of China using Partial Least Square (PLS) method. Results demonstrated that trade growth is one of the positive drivers behind the increase in CO2 emissions, and the influence of trade in underdeveloped region is greater than in the developing and highly developed region. In a recent study, Zhang  established the ARDL model to study the relationship between trade liberalization and CO2 emissions in South Korea and noted that trade has a detrimental impact on environmental quality.
On the contrary, Helpman  argued that trade helps to promote the international movement of production elements and international transfer of technical knowledge, so as to realize the efficient utilization of resources. In this way, they believed that foreign trade is conductive to control the level of environmental pollution. Furthermore, Shahbaz et al.  have undertaken an analysis on the relationship between trade openness, financial development, economic growth, carbon emissions, and coal consumption using ARDL bounds testing approach. They also confirmed that the enhancement of trade has a significant contribution to environmental improvement in South Africa. Through exploration on 25 OECD countries, Jebli et al.  investigated the correlation among carbon emissions, international trade, GDP, and renewable and nonrenewable energy consumption. Results of panel FMOLS and DOLS supported that trade is effective in mitigating carbon emissions. In another study, Zhang et al.  used the same method to study the trade-emissions nexus for ten newly industrialized countries with results that suggested that trade liberalization is essential for achieving environmental improvement.
In general, most of the previous research on the environmental consequences of FDI and foreign trade applied the mean regression. Obviously, there are still some aspects that need to be improved. First, very little research has presented the indirect impacts of FDI and foreign trade on CO2 emissions by taking economic growth as a key factor. Second, studies have failed to capture the comprehensive effects of FDI and foreign trade on the environment in pace with the changes of CO2 emissions. This paper extends earlier investigations by estimating the direct and indirect impacts of FDI and foreign trade on CO2 emissions across Chinese provinces within the panel quantile regression framework. In particular, we provide evidence that the influences of FDI and foreign trade on CO2 emissions are heterogeneous at different emissions levels.
3. Methodology and Data
3.1.1. Empirical Model Specification
In this article, the effects of FDI (Panel A) and foreign trade (Panel B) on CO2 emissions at provincial level are assessed separately for two main reasons. On the one hand, as important catalysts for economic growth in developing countries, both FDI and foreign trade reflect the openness degree of a country. On the other hand, there is a strong interaction between FDI and foreign trade . The estimated results of foreign trade can be regarded as the robustness test for FDI . To identify other socioeconomic factors affecting CO2 emissions, the paper follows the STIRPAT model proposed by Dietz and Rosa , which is defined aswhere stands for the intercept term and , , , , and represent environmental impact, population size, average affluence, technological development, and the random error term, respectively. The subscripts and denote province and year.
This study extends the STIRPAT model to reduce specification bias in econometric model estimation. First, given that the possible nonlinear correlation between economy and the environment, the squared term of GDP is included. It makes sense to verify the EKC hypothesis, because whether economic growth is achieved with the cost of rapid environmental degradation has long been the center of debate in the field of low-carbon economics. Second, due to low prices and abundant reserves, China’s energy consumption is heavily dependent on coal . As the largest energy consumers in the world, substantial energy consumption and coal-dominated energy structure are bound to cause severe environmental concerns [41, 42]. Thus, energy consumption and energy structure are introduced in the discussion. Third, recent studies, such as Xu and Lin  and Xu and Lin , considered urbanization and industrialization as the main driving forces of environmental pollution. Indeed, China is now at a stage of accelerated urbanization and industrialization, which requires huge fossil energy consumption to meet social needs and inevitably triggers ever-increasing CO2 emissions. It would be more appropriate to incorporate urbanization and industrialization into our analytical framework. All variables take a natural logarithmic form to eliminate possible heteroscedasticity and then the CO2 emissions equation can be written as follows (in this section, we only display the regression equation for FDI. Both the CO2 emissions equation and income equation for foreign trade are to replace FDI with foreign trade):where is per capita carbon emissions in tons, represents per capita of foreign direct investment (yuan), and and denote gross domestic product per capita and its square term, measured in 2000 constant yuan. indicates the total energy consumption (10,000 Standard coal), is the percentage of urban population in total population, is the percentage of industrial added value in GDP, refers to energy structure measured as the percentage of coal consumption in total energy consumption, and indicates population size (10,000 persons).
Following the augmented classical macroeconomics growth equation proposed by Solow , FDI, domestic capital stock, and population growth are controlled for in the income equation. The formulation of income equation is given as follows:where is the same as in (2), is per capita domestic capital stock (yuan), and denotes population growth.
The two-stage least squares approach developed by Kelejian  is utilized to estimate (2) and (3). The potential endogeneity problem in (3) is addressed by using an instrumental variable (IV) method. In accordance with Zhang et al. , the squared term of the fitted values of GDP per capita is chosen as the instrument variable. Additionally, the Hausman specification tests yield a p-value less than 0.05, indicating that the random effects estimates are rejected and the model with fixed effects is preferred for both (2) and (3).
For a deeper understanding of the environmental consequence of FDI, the indirect effect of FDI on CO2 emissions through its impact on per capita GDP should also be considered. Then, the total effect of FDI can be calculated by adding the direct and indirect effect as where the first item of (4) is the direct effect of FDI on CO2 emissions and the second item of the equation (i.e., the product of partial differential expression) is the indirect effect.
3.1.2. Quantile Regression
The quantile regression approach, firstly introduced by Koenker and Bassett , is a powerful tool to examine how covariates affect the location, scale, and shape at different points of response distribution. Compared with OLS regression, quantile regression proved to be more robust to the outliers and more efficient when the error term is a nonnormal and has heteroskedasticity . The quantile regression model with fixed effect can be expressed aswhere , is the unobserved effect, denotes a vector of independent variables, represents a vector of parameters to be estimated, and is the -th conditional quantile of . For a given , the parameter vector in (5) can be estimated aswhere is the quantile loss function and denotes the indicator function. Zhu et al.  noted that the estimator is robust because it splits the residuals into positives and negatives and gives weights of and . The parameter can be estimated by where is the weight that controls the relative influence of the quantiles , is the penalty function, and is the tuning parameter . For , the fixed effects estimator is obtained and the penalty term disappears, while as , the model without individual effects is obtained. Following Lamarche , the equally weighted quantiles are defined as . Accordingly, (2) can be further reconstructed to the following model:
3.2. Data Sources and Descriptions
Fossil fuel combustion is responsible for three-quarters of China’s carbon emissions . In China, there is a lack of official statistics on energy-related CO2 emissions. It is suggested that the data of CO2 emissions from different energy sources should be calculated first. Following the framework recommended by the Intergovernmental Panel on Climate Change , carbon emissions can be calculated as follows:where is CO2 emissions from fossil fuels (10,000 ton) in year , 44/12 is the conversion coefficient between C and CO2, denotes the fossil fuel consumption of measured by 104 tce (tons of coal equivalent) in the year , is the carbon emissions coefficient (t C/tce), and represents the rate of carbon oxidation of fuel type . The carbon emissions coefficient and the rate of carbon oxidation of eight fossil fuels are listed in Table 1.
Our sample dataset includes a balanced panel dataset of 30 Chinese provinces during 1997-2014. Due to data availability, Tibet, Taiwan, Hong Kong, and Macau were excluded. All data used are obtained from the China Statistical Yearbook, provincial statistical yearbooks, and China Energy Statistical Yearbook. The real GDP is standardized to be in the 2000 constant price (2000=100). Domestic capital stock is estimated by the perpetual inventory method. All variables are converted into natural logarithms except population growth rate (). Table 2 reports the details of related variables used in the model.
Note: all the sample data are obtained from the China Statistical Yearbook, provincial statistical yearbooks, and China Energy Statistical Yearbook
The results of descriptive statistics are tabulated in Table 3. The Jarque–Bera test is applied to check the normality of variables. All data series exhibit skewed and nonzero skewness, which indicates that most time-series variables reject the null hypothesis of being normally distributed, and justifies the appropriateness of applying quantile regression approach for empirical analysis.
Note: , , and denote the significance at 1%, 5%, and 10% level, respectively.
4. Empirical Results
4.1. Unit Root Test and Cointegration Test
To avoid the mistakes resulting from the pseudoregression problems, it is necessary to first check the stationarity of each variable concerned. There have been several methods for panel unit root test. Specifically, these approaches can be roughly divided into two kinds . The first is under the assumption that there is a common unit root for all panel units, including LLC (Levin-Lin-Chu) test, Breitung test, and Hadri tests. In the second category, the assumption of cross-sectional independence is relaxed and cross-sectional dependence is allowed. It primarily includes Fisher-ADF test, Fisher-PP test, and IPS (Im-Pesaran-Skin) test. Following Zhang et al. , three panel-based unit root tests, namely, Fisher-ADF, Fisher-PP, and LLC tests, are conducted to check the stationarity of each variable, and the results are depicted in Table 4. As we can see, all variables are stationary at the first-order difference, meaning that they are integrated at an order of one. Then, Kao residual cointegration test is performed to identify whether there is a long-run equilibrium relationship between variables considered. The results of the ADF stat (-6.849, P = 0.000), HAC variance = 0.0043 for panel A and ADF stat (-6.897, P = 0.000), HAC variance = 0.0043 for panel B suggest that there exist cointegration relationships between variables during the sample period.
Note: the significance probabilities for the corresponding tests are reported in parentheses.
4.2. Direct Effect Estimations Based on Panel Quantile Regression
On the basis of panel quantile regression technique, the effects of FDI and foreign trade on Chinese provincial CO2 emissions are examined in the present paper. For comparison purposes, OLS regression is firstly utilized to estimate the average impacts. The OLS estimates with pooled and one-way individual fixed effects are given in columns 1 and 2 of Table 5. As Baltagi  pointed out, the parameter estimates of the two-way fixed effect model are more reliable. Therefore, the results presented in column 3 of Table 5 are discussed below. The estimated coefficient of FDI is -0.0378 and is significant at 1% level, meaning that a 1% increase in FDI will lead to about 0.04% decrease in CO2 emissions. The result is in accordance with Zhou et al. , who noted that FDI plays a significant role in curbing carbon emissions in China. The estimated coefficient of foreign trade is -0.1042 and is significant at 1% level, showing that an increase of 1% in foreign trade will result in a reduction of about 0.10% in CO2 emissions. The finding is consistent with Destek et al. , who concluded that a 1% increase in trade will reduce CO2 emissions by 0.069%-0.097% in a panel of ten selected Central and Eastern European Countries (CEECs). However, OLS estimation results only reveal the influences of FDI and foreign trade in the center of the distribution of CO2 emissions, which may cause a loss of important information in the tail distribution.
Note: , , and denote the significance at 1%, 5%, and 10% level, respectively. Figures in parentheses are t-values.
This paper then set up two panel quantile regression models for Panels A and B to examine the direct effects of FDI and foreign trade on different quantiles of CO2 emissions in 30 Chinese provinces. The coefficients estimation for the , , , , , , , , , , and percentiles of the conditional CO2 emissions distribution is presented in Table 6. Obviously, there is a significant distinction between parameters obtained by the quantile regression and OLS regression. Figure 1 graphically displays the change of parameter estimates for main variables (i.e., FDI, foreign trade, and GDP and its square terms), ranging from 0.01 to 0.99 quantiles of the distribution with 95% confidence intervals (the shaded area).
Note: , , and denote the significance at 1%, 5%, and 10% level, respectively.
(a) OLS and quantile regression estimates
(b) OLS and quantile regression estimates
As observed from the results listed in Table 6, the direct effect of FDI on CO2 emissions is heterogeneous under low, medium, and high quantiles. At the and quantiles, the coefficients associated with FDI is negative but statistically insignificant at the 10% level. Yet, the negative coefficients of FDI become significant at other quantiles, lending strong support to the halo effect hypothesis in middle- and high-emissions provinces. These findings indicate that a 1% increase in FDI will reduce CO2 emissions by approximately 0.037%-0.073%. Furthermore, with the increase of quantiles, the degree of the impact displays an approximate increase trend, as shown in Figure 1. Therefore, it can be reasonably concluded that FDI is more conductive to mitigate CO2 emissions in high-emissions provinces. This phenomenon is attributed to important differences in environmental regulations. Provinces with high emissions tend to be more environmental conscious and enforce stricter environmental control standards. The properly designed environmental management and supervision system raises the cost of pollution-intensive products and then triggers environmental innovation. Hence, in high-emissions provinces, the influx of FDI can directly relieve the environmental pressure by transferring environmentally friendly technologies and improving energy efficiency. In Chinese provinces with rigid environmental standards, as Dean et al.  pointed out, FDI inflows into high-pollution industries can be prevented. The evidence on pollution halo hypothesis is in line with empirical results provided by Zhang and Zhou , who also reported that FDI flows are beneficial to environmental protection in China.
Similarly, the effect of foreign trade on CO2 emissions differs remarkably across quantile levels. Specifically, the coefficient of foreign trade is negative but insignificant on the left side of the CO2 emissions distribution, which means that higher level of foreign trade does not appear to decrease CO2 emissions in low-emissions provinces. However, the coefficient of foreign trade becomes statistically significant at upper quantiles (, , , , and ) and ranges from -0.0394 to -0.0684, suggesting that the expansion of foreign trade directly constrains CO2 emissions in high-emissions provinces. The negative trade–emissions relation is also documented by Dogan and Seker  for a group of the top renewable energy countries. Additionally, as CO2 emissions increase, the degree of impact of foreign trade increases gradually. That is, in high-emissions provinces, foreign trade plays more important role in reducing pollutant emissions. The reason is that the trade structure of high-emissions provinces has evolved from heavy industry to service industry and light manufacturing industry. Because of the advantages in the circulation of commodities, technology, and services, foreign trade contributes to propping up technological innovation, reducing the excess capacity, and thus improving the quality of environment. This evidence is corroborated by Wiedmann et al. , who argued that the export of industrial products results in air pollution (i.e., PM2.5), while the export of tertiary industries is an effective strategy to control pollution.
When it comes to the income-related emissions nexus, the coefficient of per capita GDP is positive, and its squared term is negative and significant at extreme low and high tails (i.e., , , and quantiles). This constitutes empirical evidence for the EKC hypothesis in the least and most polluted provinces. That is, income follows an inverse U-type with pollution emissions in this region. The result highlights that provinces with the least and most emissions may have reached the desired level of per capita income, so environmental pollution will decrease along with economic growth. Conversely, there are signs that the current economic growth mode in middle-emissions provinces seems unsustainable, since raising economic development level is accompanied by environmental deterioration. The result is contradictory to the findings of Pao et al. , who proposed that the EKC hypothesis does not support the data from China. The reason for the absence of consensus may be that ignoring the heterogeneity within provinces eliminates the existence of inverted U-shaped EKC. This finding also casts doubt on the effectiveness of policy recommendations based on mean regression results. Quantile regression technique applied in this study can completely describe the income-emissions nexus. There are significant differences in the level of environmental pollution among Chinese provinces, which further confirms the importance of discussing the EKC hypothesis under the control of distribution heterogeneity.
This paper also draws the conclusion with respect to the impacts of control variables on CO2 emissions from Table 6. First, the coefficients of energy consumption and energy structure are positive and significant for all the conditional quantiles, which indicates that massive energy consumption and coal-based energy structure will undoubtedly deteriorate the quality of environment. Second, the coefficient of urbanization suggests that the increase of urban population has a positive influence on CO2 emissions as expected. Third, the coefficient of industrialization is positive but statistically significant only at the and quantiles, which means that the industrial development is one of the causes of environmental deterioration in the least polluted provinces. Fourth, the coefficient of population size is positive and significant across quantile levels, suggesting that population growth conduces to improving the environmental quality. This may be the reason that population accumulation makes it possible to utilize energy in a relatively efficient way (e.g., central heating).
4.3. Estimation Results of CO2 Emissions Equation and Income Equation
The two-stage least square method is adopted to estimate the CO2 emissions equation and income equation represented by (2) and (3), with the CO2 emissions per capita and GDP per capita as the dependent variables, respectively. The results documented in Table 7 specify that FDI and foreign trade have negative effects on CO2 emissions. An increase of 1% in FDI and foreign trade contributes to a reduction of about 0.04% and 0.11% in CO2 emissions, respectively. Moreover, this study also presents empirical evidence on the role of FDI and foreign trade in accelerating China’s economic growth. For every 1% increase in FDI and foreign trade, GDP per capita will increase by approximately 0.27% and 0.29%. The results are broadly in accord with Lee  and Halicioglu . The former documented that FDI benefits the growth of economy of the group of 20 through direct capital financing and foreign technology transfer. The latter confirmed that foreign trade boosts economic growth in Turkey.
4.4. The Total Effects of FDI and Foreign Trade on CO2 Emissions
In this subsection, the total environmental consequences of FDI and foreign trade in 30 provinces of China are reported in Table 8. According to (4), the indirect effects of FDI and foreign trade on environment are related to income level. The different income levels are calculated based on eleven quantiles of per capita income. In addition, Figure 2 provides the graphical illustrations for the environmental impacts of FDI and foreign trade across quantile levels.
Regarding FDI, its positive indirect effect on CO2 emissions is greater than the negative direct effect and hence the total effect is positive. Our results imply that the motivation behind lowering environmental regulations to attract FDI flows has turned China into a “pollution haven.” The validation of the PHH in our study comes, however, as no surprise. It is generally acknowledged that China has become the most attractive host country of FDI due to its abundant natural resources, cheap labor cost, and lax environmental regulations. However, FDI inflows in China prefer high pollution industries such as equipment manufacturing and chemical and real estate industry and ultimately, resources are destroyed and the environment is polluted carelessly. The result is consistent with earlier studies by Kivyiro and Arminen , which revealed that the rise in FDI inflows increases pollution emissions in sub-Saharan Africa due to outdated technologies and nonmandatory regulations.
As for foreign trade, its positive indirect effect is greater than the negative direct effect within most of the income range; thus the total effect of foreign trade on CO2 emissions is also positive. It is held that increasing the liberalization of foreign trade will aggravate environmental pollution. After the accession to the World Trade Organization (WTO), China has benefited from low tariffs and nontariff barriers, and a large volume of “Made in China” products and services are exported through trade. This, however, comes at the expense of increased environmental pollutions. Under the concept of “trade causes growth,” China has neglected the adverse impact of foreign trade on environment. As a “world factory,” resource-intensive products such as steel, nonferrous metals, and petrochemical and building materials still account for a large proportion of China’s total exports. The production activities of these products inevitably consume enormous fossil energy, resulting in more severe environmental issues.
Notably, the total environmental influences of FDI and foreign trade will decrease with CO2 emissions increasing. At the quantile, as FDI or foreign trade rises by 1%, CO2 emissions will increase by approximately 0.076% or 0.085%. At the quantile, for every 1% increase in FDI or foreign trade, CO2 emissions will increase by about 0.017% and 0.044%. At the quantile, the total effects are found to be 0.006 and -0.001, respectively, which means that the direct effects of FDI and foreign trade on CO2 emissions are almost offset by the indirect effects. In view of this, FDI and foreign trade are more responsible for environmental degradation in low- and middle-emissions provinces, while, in the most polluted provinces, their environmental impacts are negligible.
5. Conclusions and Policy Implications
Using provincial-level panel data in China over the period from 1997 to 2014, this paper aims to investigate the effects of FDI and foreign trade on CO2 emissions. The panel quantile regression is applied to achieve the objectives. The results yield evidence of heterogeneity regarding the impacts of FDI and foreign trade on CO2 emissions. Specifically, the effect of FDI on CO2 emissions is negative and significant except at the and quantiles. Foreign trade has a negative and significant impact at higher quantiles, and the degree of the impact increases with the increase of quantiles. Our findings also validate the existence of EKC hypothesis in the least and most polluted provinces. However, the total environmental effects of FDI and foreign trade are determined by the indirect effects and appear positive. The above empirical findings have the following policy implications.
First, given that FDI inflows could act as a channel for environmentally friendly technologies, more stringent environmental regulations should be designed and implemented in low-emissions provinces to attract clean FDI inflows. Provinces with low- and middle-emissions are supposed to optimize the distribution structure of FDI inflows, motivating the quality-oriented mode of FDI inflows rather than a quantity-oriented mode.
Second, to upgrade trade structure in low- and middle-emissions provinces, the application of green technologies should be guaranteed in export-oriented industries. The government can take various incentives, such as tax reduction or exemption for nonpolluted industries, to encourage producers to shift toward environmentally sound industries. Again, the exports of pollution-intensive products in these areas are expected to be restricted or prohibited.
Third, to achieve the coordination between economy and carbon-cutting, it is suggested that middle-emissions provinces further adjust their economic structure and develop green economy and circular economy. Comparatively, the least and most polluted provinces should sustainably accelerate economic development, since higher income is conductive to the reduction of CO2 emission.
Finally, the key implication of the empirical findings is that emission mitigation measures should be tailored according to the pollution level in different provinces, instead of a “one-size-fits-all” approach.
Still, our study has limitations. Since environmental pollution may have potential spatial dependence characteristics among adjacent provinces in China, further analyses can examine the spatial effects of FDI and foreign trade on CO2 emissions.
The data used to support the findings of this study are available from the corresponding author upon request.
Highlights. (i) This paper employs panel quantile regression approach to investigate the effects of FDI and foreign trade on CO2 emissions in China. (ii) Heterogeneous effects of FDI and foreign trade on CO2 emissions are found. (iii) The environmental Kuznets curve (EKC) hypothesis is valid only in the least and most polluted provinces.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
This research was partly supported by the National Natural Science Foundation of China with Grant nos. 71521061, 71671062, 71471020, 51478049, and 71431008 and Major Scientific and Technological Special Program of Human Province, China, with Grant no. 2015GK1001-1.
- K. Li and B. Lin, “Economic growth model, structural transformation, and green productivity in China,” Applied Energy, vol. 187, pp. 489–500, 2017.
- S. Ren, B. Yuan, X. Ma, and X. Chen, “The impact of international trade on China's industrial carbon emissions since its entry into WTO,” Energy Policy, vol. 69, pp. 624–634, 2014.
- BP (2016) BP statistical review of world energy 2016. BP.
- Y. Liu, Y. Hao, and Y. Gao, “The environmental consequences of domestic and foreign investment: Evidence from China,” Energy Policy, vol. 108, pp. 271–280, 2017.
- G. M. Grossman and A. B. Krueger, “Economic growth and the environment,” The Quarterly Journal of Economics, vol. 110, no. 2, pp. 353–377, 1995.
- M. B. Jebli, S. B. Youssef, and I. Ozturk, “Testing environmental Kuznets curve hypothesis: The role of renewable and non-renewable energy consumption and trade in OECD countries,” Ecological Indicators, vol. 60, pp. 824–831, 2016.
- J. S. Riti, D. Song, Y. Shu, and M. Kamah, “Decoupling CO2 emission and economic growth in China: Is there consistency in estimation results in analyzing environmental Kuznets curve?” Journal of Cleaner Production, vol. 166, pp. 1448–1461, 2017.
- X. Li and X. Liu, “Foreign Direct Investment and economic growth: An increasingly endogenous relationship,” World Development, vol. 33, no. 3, pp. 393–407, 2005.
- L.-S. Lau, C.-K. Choong, and Y.-K. Eng, “Investigation of the environmental Kuznets curve for carbon emissions in Malaysia: DO foreign direct investment and trade matter?” Energy Policy, vol. 68, pp. 490–497, 2014.
- B. S. Cade and B. R. Noon, “A gentle introduction to quantile regression for ecologists,” Frontiers in Ecology and the Environment, vol. 1, no. 8, pp. 412–420, 2003.
- W.-H. You, H.-M. Zhu, K. Yu, and C. Peng, “Democracy, Financial Openness, and Global Carbon Dioxide Emissions: Heterogeneity Across Existing Emission Levels,” World Development, vol. 66, pp. 189–207, 2015.
- B. Lin and B. Xu, “Which provinces should pay more attention to CO2 emissions? Using the quantile regression to investigate China's manufacturing industry,” Journal of Cleaner Production, vol. 164, pp. 980–993, 2017.
- H. Zhu, L. Duan, Y. Guo, and K. Yu, “The effects of FDI, economic growth and energy consumption on carbon emissions in ASEAN-5: Evidence from panel quantile regression,” Economic Modelling, vol. 58, pp. 237–248, 2016.
- B. R. Copeland and M. S. Taylor, “North-South trade and the environment,” The Quarterly Journal of Economics, vol. 109, no. 3, pp. 755–787, 1994.
- M. Shahbaz, S. Nasreen, F. Abbas, and O. Anis, “Does foreign direct investment impede environmental quality in high-, middle-, and low-income countries?” Energy Economics, vol. 51, pp. 275–287, 2015.
- C. F. Tang and B. W. Tan, “The impact of energy consumption, income and foreign direct investment on carbon dioxide emissions in Vietnam,” Energy, vol. 79, no. C, pp. 447–454, 2015.
- J. Baek, “A new look at the FDI-income-energy-environment nexus: Dynamic panel data analysis of ASEAN,” Energy Policy, vol. 91, pp. 22–27, 2016.
- F. Seker, H. M. Ertugrul, and M. Cetin, “The impact of foreign direct investment on environmental quality: A bounds testing and causality analysis for Turkey,” Renewable & Sustainable Energy Reviews, vol. 52, pp. 347–356, 2015.
- S. A. Solarin, U. Al-Mulali, I. Musah, and I. Ozturk, “Investigating the pollution haven hypothesis in Ghana: An empirical investigation,” Energy, vol. 124, pp. 706–719, 2017.
- C. Sun, F. Zhang, and M. Xu, “Investigation of pollution haven hypothesis for China: An ARDL approach with breakpoint unit root tests,” Journal of Cleaner Production, vol. 161, pp. 153–164, 2017.
- S. R. Behera and D. P. Dash, “The effect of urbanization, energy consumption, and foreign direct investment on the carbon dioxide emission in the SSEA (South and Southeast Asian) region,” Renewable & Sustainable Energy Reviews, vol. 70, pp. 96–106, 2017.
- M. Salahuddin, K. Alam, I. Ozturk, and K. Sohag, “The effects of electricity consumption, economic growth, financial development and foreign direct investment on CO2 emissions in Kuwait,” Renewable & Sustainable Energy Reviews, vol. 81, pp. 2002–2010, 2018.
- X. Zhou, J. Zhang, and J. Li, “Industrial structural transformation and carbon dioxide emissions in China,” Energy Policy, vol. 57, pp. 43–51, 2013.
- M. Mert and G. Bölük, “Do foreign direct investment and renewable energy consumption affect the CO2 emissions? New evidence from a panel ARDL approach to Kyoto Annex countries,” Environmental Science and Pollution Research, vol. 23, no. 21, pp. 21669–21681, 2016.
- U. Al-mulali and C. F. Tang, “Investigating the validity of pollution haven hypothesis in the gulf cooperation council (GCC) countries,” Energy Policy, vol. 60, pp. 813–819, 2013.
- G. M. Grossman and A. B. Krueger, “Environmental Impacts of a North American Free Trade Agreement,” National Bureau of Economic Research, vol. 3914, pp. 1–57, 1991.
- M. A. Cole, “Trade, the pollution haven hypothesis and the environmental Kuznets curve: Examining the linkages,” Ecological Economics, vol. 48, no. 1, pp. 71–81, 2004.
- M. S. Hossain, “Panel estimation for CO2 emissions, energy consumption, economic growth, trade openness and urbanization of newly industrialized countries,” Energy Policy, vol. 39, no. 11, pp. 6991–6999, 2011.
- A. Sharif, S. A. Raza, I. Ozturk, and S. Afshan, “The dynamic relationship of renewable and nonrenewable energy consumption with carbon emission: A global study with the application of heterogeneous panel estimations,” Journal of Renewable Energy, vol. 133, pp. 685–691, 2019.
- A. K. Tiwari, M. Shahbaz, and Q. M. Adnan Hye, “The environmental Kuznets curve and the role of coal consumption in India: Cointegration and causality analysis in an open economy,” Renewable & Sustainable Energy Reviews, vol. 18, pp. 519–527, 2013.
- I. Ozturk and U. Al-Mulali, “Investigating the validity of the environmental Kuznets curve hypothesis in Cambodia,” Ecological Indicators, vol. 57, pp. 324–330, 2015.
- Y. Wang and T. Zhao, “Impacts of energy-related CO2 emissions: evidence from under developed, developing and highly developed regions in China,” Ecological Indicators, vol. 50, pp. 186–195, 2015.
- S. Zhang, “Is trade openness good for environment in South Korea? The role of non-fossil electricity consumption,” Environmental Science and Pollution Research, vol. 25, pp. 9510–9522, 2018.
- E. Helpman, “Explaining the structure of foreign trade: Where do we stand?” Review of World Economics, vol. 134, no. 4, pp. 587–589, 1998.
- M. Shahbaz, A. Kumar Tiwari, and M. Nasir, “The effects of financial development, economic growth, coal consumption and trade openness on CO2 emissions in South Africa,” Energy Policy, vol. 61, pp. 1452–1459, 2013.
- S. Zhang, X. Liu, and J. Bae, “Does trade openness affect CO2 emissions: evidence from ten newly industrialized countries?” Environmental Science and Pollution Research, vol. 24, no. 21, pp. 17616–17625, 2017.
- S. S. Makki and A. Somwaru, “Impact of foreign direct investment and trade on economic growth: Evidence from developing countries,” American Journal of Agricultural Economics, vol. 86, no. 3, pp. 795–801, 2004.
- Y. Hao and Y.-M. Liu, “Has the development of FDI and foreign trade contributed to China’s CO2 emissions? An empirical study with provincial panel data,” Natural Hazards, vol. 76, no. 2, pp. 1079–1091, 2015.
- T. Dietz and E. A. Rosa, “Effects of population and affluence on CO2 emissions,” Proceedings of the National Acadamy of Sciences of the United States of America, vol. 94, no. 1, pp. 175–179, 1997.
- S. Jiang, Z. Chen, L. Shan, X. Chen, and H. Wang, “Committed CO2 emissions of China's coal-fired power generators from 1993 to 2013,” Energy Policy, vol. 104, pp. 295–302, 2017.
- J. B. Ang, “CO2 emissions, energy consumption, and output in France,” Energy Policy, vol. 35, no. 10, pp. 4772–4778, 2007.
- B. Xu and B. Lin, “Assessing the development of China's new energy industry,” Energy Economics, vol. 70, pp. 116–131, 2018.
- B. Xu and B. Lin, “How industrialization and urbanization process impacts on CO2 emissions in China: evidence from nonparametric additive regression models,” Energy Economics, vol. 48, pp. 188–202, 2015.
- B. Xu and B. Lin, “A quantile regression analysis of China's provincial CO2 emissions: Where does the difference lie?” Energy Policy, vol. 98, pp. 328–342, 2016.
- R. M. Solow, “A contribution to the theory of economic growth,” The Quarterly Journal of Economics, vol. 70, no. 1, pp. 65–94, 1956.
- H. H. Kelejian, “Two-stage least squares and econometric systems linear in parameters but nonlinear in the endogenous variables,” Journal of the American Statistical Association, vol. 66, no. 334, pp. 373-374, 1971.
- Y.-J. Zhang, Y.-L. Jin, J. Chevallier, and B. Shen, “The effect of corruption on carbon dioxide emissions in APEC countries: A panel quantile regression analysis,” Technological Forecasting & Social Change, vol. 112, pp. 220–227, 2016.
- R. Koenker and G. Bassett Jr., “Regression quantiles,” Econometrica, vol. 46, no. 1, pp. 33–50, 1978.
- M. Mello and R. Perrelli, “Growth equations: A quantile regression exploration,” The Quarterly Review of Economics and Finance, vol. 43, no. 4, pp. 643–667, 2003.
- H. Zhu, Y. Guo, W. You, and Y. Xu, “The heterogeneity dependence between crude oil price changes and industry stock market returns in China: Evidence from a quantile regression approach,” Energy Economics, vol. 55, pp. 30–41, 2016.
- R. Koenker, “Quantile regression for longitudinal data,” Journal of Multivariate Analysis, vol. 91, no. 1, pp. 74–89, 2004.
- C. Lamarche, “Robust penalized quantile regression estimation for panel data,” Journal of Econometrics, vol. 157, no. 2, pp. 396–408, 2010.
- Q. Wang, S.-D. Wu, Y.-E. Zeng, and B.-W. Wu, “Exploring the relationship between urbanization, energy consumption, and CO2 emissions in different provinces of China,” Renewable & Sustainable Energy Reviews, vol. 54, pp. 1563–1579, 2016.
- Intergovernmental Panel on Climate Change (IPCC), “IPCC guidelines for national greenhouse gas inventories,” 2006, https://www.ipcc-nggip.iges.or.jp/.
- B. Lin and B. Xu, “Factors affecting CO2 emissions in China's agriculture sector: A quantile regression,” Renewable & Sustainable Energy Reviews, vol. 94, pp. 15–27, 2018.
- B. Baltagi, Econometric Analysis of Panel Data, John Wiley & Sons, Hoboken, NJ, USA, 2008.
- M. A. Destek, E. Balli, and M. Manga, “The Relationship between CO2 Emission, Energy Consumption, Urbanization and Trade Openness for Selected CEECs,” Research in World Economy, vol. 7, no. 1, 2016.
- J. M. Dean, M. E. Lovely, and H. Wang, “Are foreign investors attracted to weak environmental regulations? Evaluating the evidence from China,” Journal of Development Economics, vol. 90, no. 1, pp. 1–13, 2009.
- C. Zhang and X. Zhou, “Does foreign direct investment lead to lower CO2 emissions? Evidence from a regional analysis in China,” Renewable & Sustainable Energy Reviews, vol. 58, pp. 943–951, 2016.
- E. Dogan and F. Seker, “The influence of real output, renewable and non-renewable energy, trade and financial development on carbon emissions in the top renewable energy countries,” Renewable & Sustainable Energy Reviews, vol. 60, pp. 1074–1085, 2016.
- T. Wiedmann, M. Lenzen, K. Turner, and J. Barrett, “Examining the global environmental impact of regional consumption activities - Part 2: Review of input-output models for the assessment of environmental impacts embodied in trade,” Ecological Economics, vol. 61, no. 1, pp. 15–26, 2007.
- H. T. Pao, H. C. Fu, and C. L. Tseng, “Forecasting of CO2 emissions, energy consumption and economic growth in China using an improved grey model,” Energy, vol. 40, no. 1, pp. 400–409, 2012.
- J. W. Lee, “The contribution of foreign direct investment to clean energy use, carbon emissions and economic growth,” Energy Policy, vol. 55, pp. 483–489, 2013.
- F. Halicioglu, “An econometric study of CO2 emissions, energy consumption, income and foreign trade in Turkey,” Energy Policy, vol. 37, no. 3, pp. 1156–1164, 2009.
- P. Kivyiro and H. Arminen, “Carbon dioxide emissions, energy consumption, economic growth, and foreign direct investment: Causality analysis for Sub-Saharan Africa,” Energy, vol. 74, no. C, pp. 595–606, 2014.
Copyright © 2019 Yuan Huang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.